deepeye by deepidv
Product Hunt Shawn-Marc Melo
deepeye catches fake images, video, and voice as you browse or chat. It works in Chrome and WhatsApp with no uploads, so the check happens in the moment.
Based on reporting by Product Hunt, Shawn-Marc Melo — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
deepidv has launched deepeye, a tool aimed at spotting AI-generated and manipulated images, video, and audio while people are still browsing. It runs as a free Chrome product and also inside WhatsApp, where it checks things like profile photos, video calls, and voice notes on the spot.
The pitch is simple and a little ambitious: no uploads, no dashboard, no middlemen. deepidv says every verification is decided by proprietary AI, then validated by agentic review and sealed with cryptographic proof. Compliance sits underneath and keeps running continuously.
This isn’t being sold as a narrow detector bolted onto the side of a workflow. The company describes deepeye as part of a broader human-plus-AI verification engine and compliance suite for risk, fraud, and continuous monitoring. It also says the stack is owned end to end and works across 211+ countries.
The product comes from deepidv’s third launch on Product Hunt. The company says it’s built in San Francisco by a team that’s exited twice before, and it has also paired deepeye with Scam.AI. The launch page currently shows a 5.0 rating from 1 review and 271 followers, which is a tiny sample but a decent early signal that people are at least curious.
What stands out here is the effort to make verification disappear into normal use. Instead of asking people to stop and send files somewhere else, deepeye tries to catch the problem where it starts: in the browser, in the chat, in the call. That’s a cleaner story than the usual security theater with a dashboard attached.
My take — AI-written commentary, not fact-checked reporting
This is the right instinct: fight deepfakes where people actually encounter them, not in some solemn admin console nobody opens twice. The catch, of course, is that every company now wants to sell “cryptographic proof” and “agentic review” as if those phrases alone scare scammers. They don’t. Useful, practical, boring security usually beats grand claims, and that’s a good rule for the whole AI verification circus.
Read more about this at: Product Hunt
